FFmpeg  5.0.1
frame.h
Go to the documentation of this file.
1 /*
2  * This file is part of FFmpeg.
3  *
4  * FFmpeg is free software; you can redistribute it and/or
5  * modify it under the terms of the GNU Lesser General Public
6  * License as published by the Free Software Foundation; either
7  * version 2.1 of the License, or (at your option) any later version.
8  *
9  * FFmpeg is distributed in the hope that it will be useful,
10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12  * Lesser General Public License for more details.
13  *
14  * You should have received a copy of the GNU Lesser General Public
15  * License along with FFmpeg; if not, write to the Free Software
16  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17  */
18 
19 /**
20  * @file
21  * @ingroup lavu_frame
22  * reference-counted frame API
23  */
24 
25 #ifndef AVUTIL_FRAME_H
26 #define AVUTIL_FRAME_H
27 
28 #include <stddef.h>
29 #include <stdint.h>
30 
31 #include "avutil.h"
32 #include "buffer.h"
33 #include "dict.h"
34 #include "rational.h"
35 #include "samplefmt.h"
36 #include "pixfmt.h"
37 #include "version.h"
38 
39 
40 /**
41  * @defgroup lavu_frame AVFrame
42  * @ingroup lavu_data
43  *
44  * @{
45  * AVFrame is an abstraction for reference-counted raw multimedia data.
46  */
47 
49  /**
50  * The data is the AVPanScan struct defined in libavcodec.
51  */
53  /**
54  * ATSC A53 Part 4 Closed Captions.
55  * A53 CC bitstream is stored as uint8_t in AVFrameSideData.data.
56  * The number of bytes of CC data is AVFrameSideData.size.
57  */
59  /**
60  * Stereoscopic 3d metadata.
61  * The data is the AVStereo3D struct defined in libavutil/stereo3d.h.
62  */
64  /**
65  * The data is the AVMatrixEncoding enum defined in libavutil/channel_layout.h.
66  */
68  /**
69  * Metadata relevant to a downmix procedure.
70  * The data is the AVDownmixInfo struct defined in libavutil/downmix_info.h.
71  */
73  /**
74  * ReplayGain information in the form of the AVReplayGain struct.
75  */
77  /**
78  * This side data contains a 3x3 transformation matrix describing an affine
79  * transformation that needs to be applied to the frame for correct
80  * presentation.
81  *
82  * See libavutil/display.h for a detailed description of the data.
83  */
85  /**
86  * Active Format Description data consisting of a single byte as specified
87  * in ETSI TS 101 154 using AVActiveFormatDescription enum.
88  */
90  /**
91  * Motion vectors exported by some codecs (on demand through the export_mvs
92  * flag set in the libavcodec AVCodecContext flags2 option).
93  * The data is the AVMotionVector struct defined in
94  * libavutil/motion_vector.h.
95  */
97  /**
98  * Recommmends skipping the specified number of samples. This is exported
99  * only if the "skip_manual" AVOption is set in libavcodec.
100  * This has the same format as AV_PKT_DATA_SKIP_SAMPLES.
101  * @code
102  * u32le number of samples to skip from start of this packet
103  * u32le number of samples to skip from end of this packet
104  * u8 reason for start skip
105  * u8 reason for end skip (0=padding silence, 1=convergence)
106  * @endcode
107  */
109  /**
110  * This side data must be associated with an audio frame and corresponds to
111  * enum AVAudioServiceType defined in avcodec.h.
112  */
114  /**
115  * Mastering display metadata associated with a video frame. The payload is
116  * an AVMasteringDisplayMetadata type and contains information about the
117  * mastering display color volume.
118  */
120  /**
121  * The GOP timecode in 25 bit timecode format. Data format is 64-bit integer.
122  * This is set on the first frame of a GOP that has a temporal reference of 0.
123  */
125 
126  /**
127  * The data represents the AVSphericalMapping structure defined in
128  * libavutil/spherical.h.
129  */
131 
132  /**
133  * Content light level (based on CTA-861.3). This payload contains data in
134  * the form of the AVContentLightMetadata struct.
135  */
137 
138  /**
139  * The data contains an ICC profile as an opaque octet buffer following the
140  * format described by ISO 15076-1 with an optional name defined in the
141  * metadata key entry "name".
142  */
144 
145  /**
146  * Timecode which conforms to SMPTE ST 12-1. The data is an array of 4 uint32_t
147  * where the first uint32_t describes how many (1-3) of the other timecodes are used.
148  * The timecode format is described in the documentation of av_timecode_get_smpte_from_framenum()
149  * function in libavutil/timecode.h.
150  */
152 
153  /**
154  * HDR dynamic metadata associated with a video frame. The payload is
155  * an AVDynamicHDRPlus type and contains information for color
156  * volume transform - application 4 of SMPTE 2094-40:2016 standard.
157  */
159 
160  /**
161  * Regions Of Interest, the data is an array of AVRegionOfInterest type, the number of
162  * array element is implied by AVFrameSideData.size / AVRegionOfInterest.self_size.
163  */
165 
166  /**
167  * Encoding parameters for a video frame, as described by AVVideoEncParams.
168  */
170 
171  /**
172  * User data unregistered metadata associated with a video frame.
173  * This is the H.26[45] UDU SEI message, and shouldn't be used for any other purpose
174  * The data is stored as uint8_t in AVFrameSideData.data which is 16 bytes of
175  * uuid_iso_iec_11578 followed by AVFrameSideData.size - 16 bytes of user_data_payload_byte.
176  */
178 
179  /**
180  * Film grain parameters for a frame, described by AVFilmGrainParams.
181  * Must be present for every frame which should have film grain applied.
182  */
184 
185  /**
186  * Bounding boxes for object detection and classification,
187  * as described by AVDetectionBBoxHeader.
188  */
190 
191  /**
192  * Dolby Vision RPU raw data, suitable for passing to x265
193  * or other libraries. Array of uint8_t, with NAL emulation
194  * bytes intact.
195  */
197 
198  /**
199  * Parsed Dolby Vision metadata, suitable for passing to a software
200  * implementation. The payload is the AVDOVIMetadata struct defined in
201  * libavutil/dovi_meta.h.
202  */
204 };
205 
214 };
215 
216 
217 /**
218  * Structure to hold side data for an AVFrame.
219  *
220  * sizeof(AVFrameSideData) is not a part of the public ABI, so new fields may be added
221  * to the end with a minor bump.
222  */
223 typedef struct AVFrameSideData {
225  uint8_t *data;
226  size_t size;
230 
231 /**
232  * Structure describing a single Region Of Interest.
233  *
234  * When multiple regions are defined in a single side-data block, they
235  * should be ordered from most to least important - some encoders are only
236  * capable of supporting a limited number of distinct regions, so will have
237  * to truncate the list.
238  *
239  * When overlapping regions are defined, the first region containing a given
240  * area of the frame applies.
241  */
242 typedef struct AVRegionOfInterest {
243  /**
244  * Must be set to the size of this data structure (that is,
245  * sizeof(AVRegionOfInterest)).
246  */
247  uint32_t self_size;
248  /**
249  * Distance in pixels from the top edge of the frame to the top and
250  * bottom edges and from the left edge of the frame to the left and
251  * right edges of the rectangle defining this region of interest.
252  *
253  * The constraints on a region are encoder dependent, so the region
254  * actually affected may be slightly larger for alignment or other
255  * reasons.
256  */
257  int top;
258  int bottom;
259  int left;
260  int right;
261  /**
262  * Quantisation offset.
263  *
264  * Must be in the range -1 to +1. A value of zero indicates no quality
265  * change. A negative value asks for better quality (less quantisation),
266  * while a positive value asks for worse quality (greater quantisation).
267  *
268  * The range is calibrated so that the extreme values indicate the
269  * largest possible offset - if the rest of the frame is encoded with the
270  * worst possible quality, an offset of -1 indicates that this region
271  * should be encoded with the best possible quality anyway. Intermediate
272  * values are then interpolated in some codec-dependent way.
273  *
274  * For example, in 10-bit H.264 the quantisation parameter varies between
275  * -12 and 51. A typical qoffset value of -1/10 therefore indicates that
276  * this region should be encoded with a QP around one-tenth of the full
277  * range better than the rest of the frame. So, if most of the frame
278  * were to be encoded with a QP of around 30, this region would get a QP
279  * of around 24 (an offset of approximately -1/10 * (51 - -12) = -6.3).
280  * An extreme value of -1 would indicate that this region should be
281  * encoded with the best possible quality regardless of the treatment of
282  * the rest of the frame - that is, should be encoded at a QP of -12.
283  */
286 
287 /**
288  * This structure describes decoded (raw) audio or video data.
289  *
290  * AVFrame must be allocated using av_frame_alloc(). Note that this only
291  * allocates the AVFrame itself, the buffers for the data must be managed
292  * through other means (see below).
293  * AVFrame must be freed with av_frame_free().
294  *
295  * AVFrame is typically allocated once and then reused multiple times to hold
296  * different data (e.g. a single AVFrame to hold frames received from a
297  * decoder). In such a case, av_frame_unref() will free any references held by
298  * the frame and reset it to its original clean state before it
299  * is reused again.
300  *
301  * The data described by an AVFrame is usually reference counted through the
302  * AVBuffer API. The underlying buffer references are stored in AVFrame.buf /
303  * AVFrame.extended_buf. An AVFrame is considered to be reference counted if at
304  * least one reference is set, i.e. if AVFrame.buf[0] != NULL. In such a case,
305  * every single data plane must be contained in one of the buffers in
306  * AVFrame.buf or AVFrame.extended_buf.
307  * There may be a single buffer for all the data, or one separate buffer for
308  * each plane, or anything in between.
309  *
310  * sizeof(AVFrame) is not a part of the public ABI, so new fields may be added
311  * to the end with a minor bump.
312  *
313  * Fields can be accessed through AVOptions, the name string used, matches the
314  * C structure field name for fields accessible through AVOptions. The AVClass
315  * for AVFrame can be obtained from avcodec_get_frame_class()
316  */
317 typedef struct AVFrame {
318 #define AV_NUM_DATA_POINTERS 8
319  /**
320  * pointer to the picture/channel planes.
321  * This might be different from the first allocated byte. For video,
322  * it could even point to the end of the image data.
323  *
324  * All pointers in data and extended_data must point into one of the
325  * AVBufferRef in buf or extended_buf.
326  *
327  * Some decoders access areas outside 0,0 - width,height, please
328  * see avcodec_align_dimensions2(). Some filters and swscale can read
329  * up to 16 bytes beyond the planes, if these filters are to be used,
330  * then 16 extra bytes must be allocated.
331  *
332  * NOTE: Pointers not needed by the format MUST be set to NULL.
333  *
334  * @attention In case of video, the data[] pointers can point to the
335  * end of image data in order to reverse line order, when used in
336  * combination with negative values in the linesize[] array.
337  */
339 
340  /**
341  * For video, a positive or negative value, which is typically indicating
342  * the size in bytes of each picture line, but it can also be:
343  * - the negative byte size of lines for vertical flipping
344  * (with data[n] pointing to the end of the data
345  * - a positive or negative multiple of the byte size as for accessing
346  * even and odd fields of a frame (possibly flipped)
347  *
348  * For audio, only linesize[0] may be set. For planar audio, each channel
349  * plane must be the same size.
350  *
351  * For video the linesizes should be multiples of the CPUs alignment
352  * preference, this is 16 or 32 for modern desktop CPUs.
353  * Some code requires such alignment other code can be slower without
354  * correct alignment, for yet other it makes no difference.
355  *
356  * @note The linesize may be larger than the size of usable data -- there
357  * may be extra padding present for performance reasons.
358  *
359  * @attention In case of video, line size values can be negative to achieve
360  * a vertically inverted iteration over image lines.
361  */
363 
364  /**
365  * pointers to the data planes/channels.
366  *
367  * For video, this should simply point to data[].
368  *
369  * For planar audio, each channel has a separate data pointer, and
370  * linesize[0] contains the size of each channel buffer.
371  * For packed audio, there is just one data pointer, and linesize[0]
372  * contains the total size of the buffer for all channels.
373  *
374  * Note: Both data and extended_data should always be set in a valid frame,
375  * but for planar audio with more channels that can fit in data,
376  * extended_data must be used in order to access all channels.
377  */
378  uint8_t **extended_data;
379 
380  /**
381  * @name Video dimensions
382  * Video frames only. The coded dimensions (in pixels) of the video frame,
383  * i.e. the size of the rectangle that contains some well-defined values.
384  *
385  * @note The part of the frame intended for display/presentation is further
386  * restricted by the @ref cropping "Cropping rectangle".
387  * @{
388  */
389  int width, height;
390  /**
391  * @}
392  */
393 
394  /**
395  * number of audio samples (per channel) described by this frame
396  */
398 
399  /**
400  * format of the frame, -1 if unknown or unset
401  * Values correspond to enum AVPixelFormat for video frames,
402  * enum AVSampleFormat for audio)
403  */
404  int format;
405 
406  /**
407  * 1 -> keyframe, 0-> not
408  */
410 
411  /**
412  * Picture type of the frame.
413  */
415 
416  /**
417  * Sample aspect ratio for the video frame, 0/1 if unknown/unspecified.
418  */
420 
421  /**
422  * Presentation timestamp in time_base units (time when frame should be shown to user).
423  */
424  int64_t pts;
425 
426  /**
427  * DTS copied from the AVPacket that triggered returning this frame. (if frame threading isn't used)
428  * This is also the Presentation time of this AVFrame calculated from
429  * only AVPacket.dts values without pts values.
430  */
431  int64_t pkt_dts;
432 
433  /**
434  * Time base for the timestamps in this frame.
435  * In the future, this field may be set on frames output by decoders or
436  * filters, but its value will be by default ignored on input to encoders
437  * or filters.
438  */
440 
441  /**
442  * picture number in bitstream order
443  */
445  /**
446  * picture number in display order
447  */
449 
450  /**
451  * quality (between 1 (good) and FF_LAMBDA_MAX (bad))
452  */
453  int quality;
454 
455  /**
456  * for some private data of the user
457  */
458  void *opaque;
459 
460  /**
461  * When decoding, this signals how much the picture must be delayed.
462  * extra_delay = repeat_pict / (2*fps)
463  */
465 
466  /**
467  * The content of the picture is interlaced.
468  */
470 
471  /**
472  * If the content is interlaced, is top field displayed first.
473  */
475 
476  /**
477  * Tell user application that palette has changed from previous frame.
478  */
480 
481  /**
482  * reordered opaque 64 bits (generally an integer or a double precision float
483  * PTS but can be anything).
484  * The user sets AVCodecContext.reordered_opaque to represent the input at
485  * that time,
486  * the decoder reorders values as needed and sets AVFrame.reordered_opaque
487  * to exactly one of the values provided by the user through AVCodecContext.reordered_opaque
488  */
490 
491  /**
492  * Sample rate of the audio data.
493  */
495 
496  /**
497  * Channel layout of the audio data.
498  */
499  uint64_t channel_layout;
500 
501  /**
502  * AVBuffer references backing the data for this frame. All the pointers in
503  * data and extended_data must point inside one of the buffers in buf or
504  * extended_buf. This array must be filled contiguously -- if buf[i] is
505  * non-NULL then buf[j] must also be non-NULL for all j < i.
506  *
507  * There may be at most one AVBuffer per data plane, so for video this array
508  * always contains all the references. For planar audio with more than
509  * AV_NUM_DATA_POINTERS channels, there may be more buffers than can fit in
510  * this array. Then the extra AVBufferRef pointers are stored in the
511  * extended_buf array.
512  */
514 
515  /**
516  * For planar audio which requires more than AV_NUM_DATA_POINTERS
517  * AVBufferRef pointers, this array will hold all the references which
518  * cannot fit into AVFrame.buf.
519  *
520  * Note that this is different from AVFrame.extended_data, which always
521  * contains all the pointers. This array only contains the extra pointers,
522  * which cannot fit into AVFrame.buf.
523  *
524  * This array is always allocated using av_malloc() by whoever constructs
525  * the frame. It is freed in av_frame_unref().
526  */
528  /**
529  * Number of elements in extended_buf.
530  */
532 
535 
536 /**
537  * @defgroup lavu_frame_flags AV_FRAME_FLAGS
538  * @ingroup lavu_frame
539  * Flags describing additional frame properties.
540  *
541  * @{
542  */
543 
544 /**
545  * The frame data may be corrupted, e.g. due to decoding errors.
546  */
547 #define AV_FRAME_FLAG_CORRUPT (1 << 0)
548 /**
549  * A flag to mark the frames which need to be decoded, but shouldn't be output.
550  */
551 #define AV_FRAME_FLAG_DISCARD (1 << 2)
552 /**
553  * @}
554  */
555 
556  /**
557  * Frame flags, a combination of @ref lavu_frame_flags
558  */
559  int flags;
560 
561  /**
562  * MPEG vs JPEG YUV range.
563  * - encoding: Set by user
564  * - decoding: Set by libavcodec
565  */
567 
569 
571 
572  /**
573  * YUV colorspace type.
574  * - encoding: Set by user
575  * - decoding: Set by libavcodec
576  */
578 
580 
581  /**
582  * frame timestamp estimated using various heuristics, in stream time base
583  * - encoding: unused
584  * - decoding: set by libavcodec, read by user.
585  */
587 
588  /**
589  * reordered pos from the last AVPacket that has been input into the decoder
590  * - encoding: unused
591  * - decoding: Read by user.
592  */
593  int64_t pkt_pos;
594 
595  /**
596  * duration of the corresponding packet, expressed in
597  * AVStream->time_base units, 0 if unknown.
598  * - encoding: unused
599  * - decoding: Read by user.
600  */
601  int64_t pkt_duration;
602 
603  /**
604  * metadata.
605  * - encoding: Set by user.
606  * - decoding: Set by libavcodec.
607  */
609 
610  /**
611  * decode error flags of the frame, set to a combination of
612  * FF_DECODE_ERROR_xxx flags if the decoder produced a frame, but there
613  * were errors during the decoding.
614  * - encoding: unused
615  * - decoding: set by libavcodec, read by user.
616  */
618 #define FF_DECODE_ERROR_INVALID_BITSTREAM 1
619 #define FF_DECODE_ERROR_MISSING_REFERENCE 2
620 #define FF_DECODE_ERROR_CONCEALMENT_ACTIVE 4
621 #define FF_DECODE_ERROR_DECODE_SLICES 8
622 
623  /**
624  * number of audio channels, only used for audio.
625  * - encoding: unused
626  * - decoding: Read by user.
627  */
628  int channels;
629 
630  /**
631  * size of the corresponding packet containing the compressed
632  * frame.
633  * It is set to a negative value if unknown.
634  * - encoding: unused
635  * - decoding: set by libavcodec, read by user.
636  */
637  int pkt_size;
638 
639  /**
640  * For hwaccel-format frames, this should be a reference to the
641  * AVHWFramesContext describing the frame.
642  */
644 
645  /**
646  * AVBufferRef for free use by the API user. FFmpeg will never check the
647  * contents of the buffer ref. FFmpeg calls av_buffer_unref() on it when
648  * the frame is unreferenced. av_frame_copy_props() calls create a new
649  * reference with av_buffer_ref() for the target frame's opaque_ref field.
650  *
651  * This is unrelated to the opaque field, although it serves a similar
652  * purpose.
653  */
655 
656  /**
657  * @anchor cropping
658  * @name Cropping
659  * Video frames only. The number of pixels to discard from the the
660  * top/bottom/left/right border of the frame to obtain the sub-rectangle of
661  * the frame intended for presentation.
662  * @{
663  */
664  size_t crop_top;
665  size_t crop_bottom;
666  size_t crop_left;
667  size_t crop_right;
668  /**
669  * @}
670  */
671 
672  /**
673  * AVBufferRef for internal use by a single libav* library.
674  * Must not be used to transfer data between libraries.
675  * Has to be NULL when ownership of the frame leaves the respective library.
676  *
677  * Code outside the FFmpeg libs should never check or change the contents of the buffer ref.
678  *
679  * FFmpeg calls av_buffer_unref() on it when the frame is unreferenced.
680  * av_frame_copy_props() calls create a new reference with av_buffer_ref()
681  * for the target frame's private_ref field.
682  */
684 } AVFrame;
685 
686 
687 #if FF_API_COLORSPACE_NAME
688 /**
689  * Get the name of a colorspace.
690  * @return a static string identifying the colorspace; can be NULL.
691  * @deprecated use av_color_space_name()
692  */
694 const char *av_get_colorspace_name(enum AVColorSpace val);
695 #endif
696 /**
697  * Allocate an AVFrame and set its fields to default values. The resulting
698  * struct must be freed using av_frame_free().
699  *
700  * @return An AVFrame filled with default values or NULL on failure.
701  *
702  * @note this only allocates the AVFrame itself, not the data buffers. Those
703  * must be allocated through other means, e.g. with av_frame_get_buffer() or
704  * manually.
705  */
707 
708 /**
709  * Free the frame and any dynamically allocated objects in it,
710  * e.g. extended_data. If the frame is reference counted, it will be
711  * unreferenced first.
712  *
713  * @param frame frame to be freed. The pointer will be set to NULL.
714  */
716 
717 /**
718  * Set up a new reference to the data described by the source frame.
719  *
720  * Copy frame properties from src to dst and create a new reference for each
721  * AVBufferRef from src.
722  *
723  * If src is not reference counted, new buffers are allocated and the data is
724  * copied.
725  *
726  * @warning: dst MUST have been either unreferenced with av_frame_unref(dst),
727  * or newly allocated with av_frame_alloc() before calling this
728  * function, or undefined behavior will occur.
729  *
730  * @return 0 on success, a negative AVERROR on error
731  */
732 int av_frame_ref(AVFrame *dst, const AVFrame *src);
733 
734 /**
735  * Create a new frame that references the same data as src.
736  *
737  * This is a shortcut for av_frame_alloc()+av_frame_ref().
738  *
739  * @return newly created AVFrame on success, NULL on error.
740  */
742 
743 /**
744  * Unreference all the buffers referenced by frame and reset the frame fields.
745  */
747 
748 /**
749  * Move everything contained in src to dst and reset src.
750  *
751  * @warning: dst is not unreferenced, but directly overwritten without reading
752  * or deallocating its contents. Call av_frame_unref(dst) manually
753  * before calling this function to ensure that no memory is leaked.
754  */
756 
757 /**
758  * Allocate new buffer(s) for audio or video data.
759  *
760  * The following fields must be set on frame before calling this function:
761  * - format (pixel format for video, sample format for audio)
762  * - width and height for video
763  * - nb_samples and channel_layout for audio
764  *
765  * This function will fill AVFrame.data and AVFrame.buf arrays and, if
766  * necessary, allocate and fill AVFrame.extended_data and AVFrame.extended_buf.
767  * For planar formats, one buffer will be allocated for each plane.
768  *
769  * @warning: if frame already has been allocated, calling this function will
770  * leak memory. In addition, undefined behavior can occur in certain
771  * cases.
772  *
773  * @param frame frame in which to store the new buffers.
774  * @param align Required buffer size alignment. If equal to 0, alignment will be
775  * chosen automatically for the current CPU. It is highly
776  * recommended to pass 0 here unless you know what you are doing.
777  *
778  * @return 0 on success, a negative AVERROR on error.
779  */
781 
782 /**
783  * Check if the frame data is writable.
784  *
785  * @return A positive value if the frame data is writable (which is true if and
786  * only if each of the underlying buffers has only one reference, namely the one
787  * stored in this frame). Return 0 otherwise.
788  *
789  * If 1 is returned the answer is valid until av_buffer_ref() is called on any
790  * of the underlying AVBufferRefs (e.g. through av_frame_ref() or directly).
791  *
792  * @see av_frame_make_writable(), av_buffer_is_writable()
793  */
795 
796 /**
797  * Ensure that the frame data is writable, avoiding data copy if possible.
798  *
799  * Do nothing if the frame is writable, allocate new buffers and copy the data
800  * if it is not.
801  *
802  * @return 0 on success, a negative AVERROR on error.
803  *
804  * @see av_frame_is_writable(), av_buffer_is_writable(),
805  * av_buffer_make_writable()
806  */
808 
809 /**
810  * Copy the frame data from src to dst.
811  *
812  * This function does not allocate anything, dst must be already initialized and
813  * allocated with the same parameters as src.
814  *
815  * This function only copies the frame data (i.e. the contents of the data /
816  * extended data arrays), not any other properties.
817  *
818  * @return >= 0 on success, a negative AVERROR on error.
819  */
820 int av_frame_copy(AVFrame *dst, const AVFrame *src);
821 
822 /**
823  * Copy only "metadata" fields from src to dst.
824  *
825  * Metadata for the purpose of this function are those fields that do not affect
826  * the data layout in the buffers. E.g. pts, sample rate (for audio) or sample
827  * aspect ratio (for video), but not width/height or channel layout.
828  * Side data is also copied.
829  */
830 int av_frame_copy_props(AVFrame *dst, const AVFrame *src);
831 
832 /**
833  * Get the buffer reference a given data plane is stored in.
834  *
835  * @param plane index of the data plane of interest in frame->extended_data.
836  *
837  * @return the buffer reference that contains the plane or NULL if the input
838  * frame is not valid.
839  */
841 
842 /**
843  * Add a new side data to a frame.
844  *
845  * @param frame a frame to which the side data should be added
846  * @param type type of the added side data
847  * @param size size of the side data
848  *
849  * @return newly added side data on success, NULL on error
850  */
852  enum AVFrameSideDataType type,
853  size_t size);
854 
855 /**
856  * Add a new side data to a frame from an existing AVBufferRef
857  *
858  * @param frame a frame to which the side data should be added
859  * @param type the type of the added side data
860  * @param buf an AVBufferRef to add as side data. The ownership of
861  * the reference is transferred to the frame.
862  *
863  * @return newly added side data on success, NULL on error. On failure
864  * the frame is unchanged and the AVBufferRef remains owned by
865  * the caller.
866  */
868  enum AVFrameSideDataType type,
869  AVBufferRef *buf);
870 
871 /**
872  * @return a pointer to the side data of a given type on success, NULL if there
873  * is no side data with such type in this frame.
874  */
876  enum AVFrameSideDataType type);
877 
878 /**
879  * Remove and free all side data instances of the given type.
880  */
882 
883 
884 /**
885  * Flags for frame cropping.
886  */
887 enum {
888  /**
889  * Apply the maximum possible cropping, even if it requires setting the
890  * AVFrame.data[] entries to unaligned pointers. Passing unaligned data
891  * to FFmpeg API is generally not allowed, and causes undefined behavior
892  * (such as crashes). You can pass unaligned data only to FFmpeg APIs that
893  * are explicitly documented to accept it. Use this flag only if you
894  * absolutely know what you are doing.
895  */
897 };
898 
899 /**
900  * Crop the given video AVFrame according to its crop_left/crop_top/crop_right/
901  * crop_bottom fields. If cropping is successful, the function will adjust the
902  * data pointers and the width/height fields, and set the crop fields to 0.
903  *
904  * In all cases, the cropping boundaries will be rounded to the inherent
905  * alignment of the pixel format. In some cases, such as for opaque hwaccel
906  * formats, the left/top cropping is ignored. The crop fields are set to 0 even
907  * if the cropping was rounded or ignored.
908  *
909  * @param frame the frame which should be cropped
910  * @param flags Some combination of AV_FRAME_CROP_* flags, or 0.
911  *
912  * @return >= 0 on success, a negative AVERROR on error. If the cropping fields
913  * were invalid, AVERROR(ERANGE) is returned, and nothing is changed.
914  */
916 
917 /**
918  * @return a string identifying the side data type
919  */
921 
922 /**
923  * @}
924  */
925 
926 #endif /* AVUTIL_FRAME_H */
#define attribute_deprecated
Definition: attributes.h:100
Convenience header that includes libavutil's core.
refcounted data buffer API
static AVFrame * frame
Public dictionary API.
#define AV_NUM_DATA_POINTERS
Definition: frame.h:318
struct AVDictionary AVDictionary
Definition: dict.h:84
void av_frame_unref(AVFrame *frame)
Unreference all the buffers referenced by frame and reset the frame fields.
void av_frame_remove_side_data(AVFrame *frame, enum AVFrameSideDataType type)
Remove and free all side data instances of the given type.
int av_frame_is_writable(AVFrame *frame)
Check if the frame data is writable.
AVFrame * av_frame_clone(const AVFrame *src)
Create a new frame that references the same data as src.
const char * av_frame_side_data_name(enum AVFrameSideDataType type)
AVActiveFormatDescription
Definition: frame.h:206
int av_frame_get_buffer(AVFrame *frame, int align)
Allocate new buffer(s) for audio or video data.
void av_frame_move_ref(AVFrame *dst, AVFrame *src)
Move everything contained in src to dst and reset src.
AVBufferRef * av_frame_get_plane_buffer(AVFrame *frame, int plane)
Get the buffer reference a given data plane is stored in.
int av_frame_ref(AVFrame *dst, const AVFrame *src)
Set up a new reference to the data described by the source frame.
void av_frame_free(AVFrame **frame)
Free the frame and any dynamically allocated objects in it, e.g.
int av_frame_copy_props(AVFrame *dst, const AVFrame *src)
Copy only "metadata" fields from src to dst.
AVFrameSideData * av_frame_new_side_data_from_buf(AVFrame *frame, enum AVFrameSideDataType type, AVBufferRef *buf)
Add a new side data to a frame from an existing AVBufferRef.
AVFrame * av_frame_alloc(void)
Allocate an AVFrame and set its fields to default values.
attribute_deprecated const char * av_get_colorspace_name(enum AVColorSpace val)
Get the name of a colorspace.
int av_frame_make_writable(AVFrame *frame)
Ensure that the frame data is writable, avoiding data copy if possible.
AVFrameSideData * av_frame_get_side_data(const AVFrame *frame, enum AVFrameSideDataType type)
AVFrameSideDataType
Definition: frame.h:48
AVFrameSideData * av_frame_new_side_data(AVFrame *frame, enum AVFrameSideDataType type, size_t size)
Add a new side data to a frame.
int av_frame_apply_cropping(AVFrame *frame, int flags)
Crop the given video AVFrame according to its crop_left/crop_top/crop_right/ crop_bottom fields.
int av_frame_copy(AVFrame *dst, const AVFrame *src)
Copy the frame data from src to dst.
@ AV_AFD_14_9
Definition: frame.h:210
@ AV_AFD_SAME
Definition: frame.h:207
@ AV_AFD_16_9_SP_14_9
Definition: frame.h:212
@ AV_AFD_4_3_SP_14_9
Definition: frame.h:211
@ AV_AFD_SP_4_3
Definition: frame.h:213
@ AV_AFD_4_3
Definition: frame.h:208
@ AV_AFD_16_9
Definition: frame.h:209
@ AV_FRAME_CROP_UNALIGNED
Apply the maximum possible cropping, even if it requires setting the AVFrame.data[] entries to unalig...
Definition: frame.h:896
@ AV_FRAME_DATA_GOP_TIMECODE
The GOP timecode in 25 bit timecode format.
Definition: frame.h:124
@ AV_FRAME_DATA_DOVI_METADATA
Parsed Dolby Vision metadata, suitable for passing to a software implementation.
Definition: frame.h:203
@ AV_FRAME_DATA_SPHERICAL
The data represents the AVSphericalMapping structure defined in libavutil/spherical....
Definition: frame.h:130
@ AV_FRAME_DATA_CONTENT_LIGHT_LEVEL
Content light level (based on CTA-861.3).
Definition: frame.h:136
@ AV_FRAME_DATA_PANSCAN
The data is the AVPanScan struct defined in libavcodec.
Definition: frame.h:52
@ AV_FRAME_DATA_DISPLAYMATRIX
This side data contains a 3x3 transformation matrix describing an affine transformation that needs to...
Definition: frame.h:84
@ AV_FRAME_DATA_A53_CC
ATSC A53 Part 4 Closed Captions.
Definition: frame.h:58
@ AV_FRAME_DATA_SEI_UNREGISTERED
User data unregistered metadata associated with a video frame.
Definition: frame.h:177
@ AV_FRAME_DATA_AUDIO_SERVICE_TYPE
This side data must be associated with an audio frame and corresponds to enum AVAudioServiceType defi...
Definition: frame.h:113
@ AV_FRAME_DATA_DYNAMIC_HDR_PLUS
HDR dynamic metadata associated with a video frame.
Definition: frame.h:158
@ AV_FRAME_DATA_REPLAYGAIN
ReplayGain information in the form of the AVReplayGain struct.
Definition: frame.h:76
@ AV_FRAME_DATA_SKIP_SAMPLES
Recommmends skipping the specified number of samples.
Definition: frame.h:108
@ AV_FRAME_DATA_MASTERING_DISPLAY_METADATA
Mastering display metadata associated with a video frame.
Definition: frame.h:119
@ AV_FRAME_DATA_DOWNMIX_INFO
Metadata relevant to a downmix procedure.
Definition: frame.h:72
@ AV_FRAME_DATA_FILM_GRAIN_PARAMS
Film grain parameters for a frame, described by AVFilmGrainParams.
Definition: frame.h:183
@ AV_FRAME_DATA_VIDEO_ENC_PARAMS
Encoding parameters for a video frame, as described by AVVideoEncParams.
Definition: frame.h:169
@ AV_FRAME_DATA_DETECTION_BBOXES
Bounding boxes for object detection and classification, as described by AVDetectionBBoxHeader.
Definition: frame.h:189
@ AV_FRAME_DATA_ICC_PROFILE
The data contains an ICC profile as an opaque octet buffer following the format described by ISO 1507...
Definition: frame.h:143
@ AV_FRAME_DATA_AFD
Active Format Description data consisting of a single byte as specified in ETSI TS 101 154 using AVAc...
Definition: frame.h:89
@ AV_FRAME_DATA_S12M_TIMECODE
Timecode which conforms to SMPTE ST 12-1.
Definition: frame.h:151
@ AV_FRAME_DATA_MATRIXENCODING
The data is the AVMatrixEncoding enum defined in libavutil/channel_layout.h.
Definition: frame.h:67
@ AV_FRAME_DATA_DOVI_RPU_BUFFER
Dolby Vision RPU raw data, suitable for passing to x265 or other libraries.
Definition: frame.h:196
@ AV_FRAME_DATA_STEREO3D
Stereoscopic 3d metadata.
Definition: frame.h:63
@ AV_FRAME_DATA_MOTION_VECTORS
Motion vectors exported by some codecs (on demand through the export_mvs flag set in the libavcodec A...
Definition: frame.h:96
@ AV_FRAME_DATA_REGIONS_OF_INTEREST
Regions Of Interest, the data is an array of AVRegionOfInterest type, the number of array element is ...
Definition: frame.h:164
AVPictureType
Definition: avutil.h:272
Libavutil version macros.
pixel format definitions
AVChromaLocation
Location of chroma samples.
Definition: pixfmt.h:616
AVColorRange
Visual content value range.
Definition: pixfmt.h:562
AVColorPrimaries
Chromaticity coordinates of the source primaries.
Definition: pixfmt.h:469
AVColorTransferCharacteristic
Color Transfer Characteristic.
Definition: pixfmt.h:494
AVColorSpace
YUV colorspace type.
Definition: pixfmt.h:523
Utilties for rational number calculation.
A reference to a data buffer.
Definition: buffer.h:82
Structure to hold side data for an AVFrame.
Definition: frame.h:223
enum AVFrameSideDataType type
Definition: frame.h:224
AVDictionary * metadata
Definition: frame.h:227
size_t size
Definition: frame.h:226
uint8_t * data
Definition: frame.h:225
AVBufferRef * buf
Definition: frame.h:228
This structure describes decoded (raw) audio or video data.
Definition: frame.h:317
int coded_picture_number
picture number in bitstream order
Definition: frame.h:444
int nb_samples
number of audio samples (per channel) described by this frame
Definition: frame.h:397
int64_t pts
Presentation timestamp in time_base units (time when frame should be shown to user).
Definition: frame.h:424
int64_t best_effort_timestamp
frame timestamp estimated using various heuristics, in stream time base
Definition: frame.h:586
int display_picture_number
picture number in display order
Definition: frame.h:448
size_t crop_right
Definition: frame.h:667
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
Definition: frame.h:338
enum AVChromaLocation chroma_location
Definition: frame.h:579
int width
Definition: frame.h:389
AVBufferRef ** extended_buf
For planar audio which requires more than AV_NUM_DATA_POINTERS AVBufferRef pointers,...
Definition: frame.h:527
AVBufferRef * hw_frames_ctx
For hwaccel-format frames, this should be a reference to the AVHWFramesContext describing the frame.
Definition: frame.h:643
int key_frame
1 -> keyframe, 0-> not
Definition: frame.h:409
AVRational time_base
Time base for the timestamps in this frame.
Definition: frame.h:439
int64_t pkt_duration
duration of the corresponding packet, expressed in AVStream->time_base units, 0 if unknown.
Definition: frame.h:601
void * opaque
for some private data of the user
Definition: frame.h:458
int64_t pkt_pos
reordered pos from the last AVPacket that has been input into the decoder
Definition: frame.h:593
int pkt_size
size of the corresponding packet containing the compressed frame.
Definition: frame.h:637
int decode_error_flags
decode error flags of the frame, set to a combination of FF_DECODE_ERROR_xxx flags if the decoder pro...
Definition: frame.h:617
int height
Definition: frame.h:389
AVFrameSideData ** side_data
Definition: frame.h:533
int flags
Frame flags, a combination of AV_FRAME_FLAGS.
Definition: frame.h:559
AVBufferRef * buf[AV_NUM_DATA_POINTERS]
AVBuffer references backing the data for this frame.
Definition: frame.h:513
int channels
number of audio channels, only used for audio.
Definition: frame.h:628
enum AVColorPrimaries color_primaries
Definition: frame.h:568
AVDictionary * metadata
metadata.
Definition: frame.h:608
uint64_t channel_layout
Channel layout of the audio data.
Definition: frame.h:499
int palette_has_changed
Tell user application that palette has changed from previous frame.
Definition: frame.h:479
AVRational sample_aspect_ratio
Sample aspect ratio for the video frame, 0/1 if unknown/unspecified.
Definition: frame.h:419
AVBufferRef * private_ref
AVBufferRef for internal use by a single libav* library.
Definition: frame.h:683
int top_field_first
If the content is interlaced, is top field displayed first.
Definition: frame.h:474
int quality
quality (between 1 (good) and FF_LAMBDA_MAX (bad))
Definition: frame.h:453
int interlaced_frame
The content of the picture is interlaced.
Definition: frame.h:469
AVBufferRef * opaque_ref
AVBufferRef for free use by the API user.
Definition: frame.h:654
size_t crop_top
Definition: frame.h:664
enum AVColorRange color_range
MPEG vs JPEG YUV range.
Definition: frame.h:566
int nb_side_data
Definition: frame.h:534
enum AVColorSpace colorspace
YUV colorspace type.
Definition: frame.h:577
int nb_extended_buf
Number of elements in extended_buf.
Definition: frame.h:531
int64_t pkt_dts
DTS copied from the AVPacket that triggered returning this frame.
Definition: frame.h:431
int linesize[AV_NUM_DATA_POINTERS]
For video, a positive or negative value, which is typically indicating the size in bytes of each pict...
Definition: frame.h:362
size_t crop_left
Definition: frame.h:666
enum AVColorTransferCharacteristic color_trc
Definition: frame.h:570
int64_t reordered_opaque
reordered opaque 64 bits (generally an integer or a double precision float PTS but can be anything).
Definition: frame.h:489
int sample_rate
Sample rate of the audio data.
Definition: frame.h:494
size_t crop_bottom
Definition: frame.h:665
int format
format of the frame, -1 if unknown or unset Values correspond to enum AVPixelFormat for video frames,...
Definition: frame.h:404
int repeat_pict
When decoding, this signals how much the picture must be delayed.
Definition: frame.h:464
enum AVPictureType pict_type
Picture type of the frame.
Definition: frame.h:414
uint8_t ** extended_data
pointers to the data planes/channels.
Definition: frame.h:378
Rational number (pair of numerator and denominator).
Definition: rational.h:58
Structure describing a single Region Of Interest.
Definition: frame.h:242
uint32_t self_size
Must be set to the size of this data structure (that is, sizeof(AVRegionOfInterest)).
Definition: frame.h:247
AVRational qoffset
Quantisation offset.
Definition: frame.h:284
int top
Distance in pixels from the top edge of the frame to the top and bottom edges and from the left edge ...
Definition: frame.h:257